Johnny Wong, PhD; Alana Slomovic, BSc; George Ibrahim, PhD; Ivan Radovanovic, PhD; Michael Tymianski, MD, PhD

Transcription

1 Microsurgery for ARUBA Trial (A Randomized Trial of Unruptured Brain Arteriovenous Malformation) Eligible Unruptured Brain Arteriovenous Malformations Johnny Wong, PhD; Alana Slomovic, BSc; George Ibrahim, PhD; Ivan Radovanovic, PhD; Michael Tymianski, MD, PhD Background and Purpose The management of unruptured brain arteriovenous malformations (ubavms) remains controversial despite ARUBA trial (A Randomized Trial of Unruptured Brain Arteriovenous Malformation), a controlled trial that suggested superiority of conservative management over intervention. However, microsurgery occurred in only 14.9% of ARUBA intervention cases, raising concerns about the study s generalizability. Our purpose was to evaluate whether, in a larger ARUBA-eligible ubavm population, microsurgery produces acceptable outcomes. Methods Demographic data, AVM characteristics, and treatment outcomes were evaluated in 155 ARUBA-eligible bavms treated with microsurgery between 1994 and Outcomes were rates of early disabling deficits and permanent disabling deficits with modified Rankin Scale score 3 or any permanent neurological deficits with modified Rankin Scale score 1. Covariates associated with outcomes were determined by regression analysis. Results Of 977 AVM patients, 155 ARUBA-eligible patients had microsurgical resection (71.6% surgery only and 25.2% with preoperative embolization). Mean follow-up was 36.1 months. Complete obliteration was achieved in 94.2% after initial surgery and 98.1% on final angiography. Early disabling deficits and permanent disabling deficits occurred in 12.3% and 4.5%, respectively, whereas any permanent neurological deficit (modified Rankin Scale score 1) occurred in 16.1%. Among ubavm of Spetzler Martin grades 1 and 2, complete obliteration occurred in 99.2%, with early disabling deficits and permanent disabling deficits occurring in 9.3% and 3.4%, respectively. Major bleeding was the only significant predictor of early disabling deficits on multivariate analysis (P<0.001). Conclusions Microsurgery in this cohort produced less disabling deficits than ARUBA with similar morbidity and AVM obliteration as other cohort series. This disparity between our results and ARUBA suggests that future controlled trials should focus on the safety and efficacy of microsurgery with or without adjunctive embolization in carefully selected ubavm patients. (Stroke. 2017;48: DOI: /STROKEAHA ) Key Words: angiography arteriovenous malformations microsurgery radiosurgery regression analysis Unruptured brain arteriovenous malformations (ubavms) are thought to be associated with an annual risk of a brain hemorrhage of 1% to 4%. 1,2 Management has been based on the perception that obliteration of ubavms is beneficial by eliminating the risk of future hemorrhage. However, this is controversial and more so after ARUBA trial (A Randomized Trial of Unruptured Brain Arteriovenous Malformation). 2 This trial compared conservative management with interventional treatment that included any one or more of microsurgery, embolization, or radiosurgery for ubavms. 2 It attempted to evaluate the balance between the natural history risk of hemorrhage from ubavms against that inherent in intervention using any of these treatment modalities. ARUBA was terminated early by the National Institutes of Health because of the superiority of medical management over interventional therapy. 2 The primary outcome (composite event of death and symptomatic stroke) was identified in 10.1% of the conservative management group when compared with 30.7% in the interventional treatment group at a mean follow-up of 33 months. ARUBA, therefore, validated the natural history of ubavms and seemingly justified a shift to conservative management. However, a major concern with ARUBA is the heterogeneity of interventional modalities used and the lack of consistent treatment assignment criteria. Of the 114 patients in the interventional group, only 14.9% (17 patients) received microsurgery, either alone (5 patients) or in combination with endovascular or radiosurgery. Because of the small surgical cohort, it is unclear whether the overall conclusions of ARUBA should be generalized to microsurgical treatment. One perspective is that microsurgery offers the benefit of immediate cure, eliminating further life-time risk of hemorrhage, and may be safely performed in properly selected cases. 3,4 Adopting the Received July 7, 2016; final revision received September 12, 2016; accepted October 7, From the Division of Neurosurgery, Toronto Western Hospital, University Health Network, Ontario, Canada (J.W., A.S., G.I., I.R., M.T.) and Department of Surgery, University of Toronto, Ontario, Canada (J.W., G.I., I.R., M.T.). Correspondence to Michael Tymianski, MD, PhD, Division of Neurosurgery, Toronto Western Hospital, 4W-435, 399 Bathurst St, Toronto, Ontario M5T 2S8, Canada. mike.tymianski@uhn.ca 2016 American Heart Association, Inc. Stroke is available at DOI: /STROKEAHA

2 Wong et al Microsurgery for Unruptured Brain AVMs 137 overarching conclusion of ARUBA could, therefore, unnecessarily deprive certain ubavm patients of a beneficial therapy. Our purpose was to evaluate whether there exists a subset of ARUBA-eligible patients selected by clinical judgment and AVM characteristics that can be safely treated by microsurgical resection. A secondary goal was to identify any prognostic factors for outcome in this patient cohort. Methods Data Collection This study was approved by the Research Ethics Board at University Health Network and was conducted in accordance with the institutional ethics guidelines. The University of Toronto Brain AVM study group database is a prospectively collected database containing demographic, clinical, and radiological information. To identify patients with ubavms treated by microsurgical resection at Toronto Western Hospital between 1994 and 2014, the database was interrogated using the search terms: arteriovenous malformation and AVM. All patients with an unruptured intracerebral AVM and treated with microsurgery were included in the study. Exclusion criteria included evidence of previous intracranial hemorrhage on computed tomography or magnetic resonance imaging, diagnosis of other vascular malformations (eg, cavernous malformations, facial, or body AVM), or any treatment without microsurgical resection of AVM. All AVM cases were diagnosed based on magnetic resonance imaging, computed tomography angiogram, or digital subtraction angiography (DSA). Management of each individual ubavm was discussed at multidisciplinary conference, in which multimodality treatment strategies, including microsurgery, embolization or radiosurgery alone, or in combination, were considered. Suitability for microsurgery and need for preoperative embolization were determined by multidisciplinary consensus, based on Spetzler Martin (SM) grade, AVM location, angioarchitecture, including the presence of high-risk features, high-flow shunting, or associated aneurysms. Operations were performed by 4 neurosurgeons. Postoperative DSA was performed within the first week of resection to verify bavm resection. Clinical follow-up was performed at 6 weeks and 3 to 6 months post hospital discharge and at annual intervals thereafter. The database and clinical records were retrieved and analyzed retrospectively by 2 individuals who were not directly involved in the care of the patients (J.W. and A.S.). Study Variables Study variables included patient demographic data, clinical presentation (seizures, headaches, neurological deficits, bruits, or asymptomatic if not relevant to the AVM), AVM characteristics (SM grading based on size, eloquence, deep venous drainage, as well as location, and associated aneurysms), pre- and post-treatment functional outcomes (modified Rankin Scale scores [mrs]), and treatment outcomes (AVM obliteration rates and new neurological deficits). AVM obliteration was confirmed by DSA. Neurological deficits were defined as early disabling deficit (EDD) if mrs 3 within 7 days of surgery and permanent disabling deficit (PDD) if mrs 3 at the last clinical follow-up. Other complications were recorded, including major bleeding (defined as >1000 ml intraoperative blood loss, or transfusion requirement for 2 U of whole or packed red cells 5 ), postoperative hematomas requiring evacuation, and wound infection. bavm were dichotomized into low grade (SM grades 1 and 2) and high grade (SM grades 3 5). Statistical Analysis Where stated, Fisher exact tests were used for categorical variables, and 2-sided t tests for continuous variables. A univariate logistic regression analysis was performed using the dichotomized outcome variable as the dependent variable to determine whether covariates were associated with the primary outcomes (permanent neurological deficit, EDD, and PDD). Significant covariates on univariate analysis (P<0.05) were entered into a multivariate logistic regression. Twoand three-way interactions were tested. Analyses were performed using R statistics software (version 3.2.1). Results Patient Demographics and AVM Characteristics From 1994 to 2014, 977 bavm patients were treated of which 528 were ubavms. Surgery occurred in 168 patients, but 10 patients were excluded because of inadequate data because the surgeries were performed before their referral to our institution, whereas in 3 patients the surgery consisted only of a decompressive craniectomy to treat complications of AVM embolization, leaving 155 patients (Figure 1). Mean followup was 36.1 months (range months). Baseline demographics, clinical presentation, and AVM characteristics are presented in Tables 1 and 2. There were 88 females (57%) and 67 males (43%); mean age at presentation was 38.5 years (range years), including 8 patients aged <18 years. The most common clinical presentations were symptomatic seizures (47.7%), focal neurological deficits (8.4%), or asymptomatic/incidental radiological findings (43.9%). Multiple AVMs occurred in 3 patients (1.9%). The most common AVM locations were frontal (29.6%), parietal (27.7%), and temporal (25.2%). SM grading was distributed accordingly: 52 grade 1 (33.5%), 66 grade 2 (42.6%), 30 grade 3 (19.4%), and 7 grade 4 bavms (4.5%). No SM grade 5 bavms were resected in this series. Intracranial aneurysms were present in 15.5% of patients. Treatment and Outcomes Treatment modalities, outcomes, and complications are presented in Figure 2 and Table 3. Microsurgery alone was performed in 107 cases (69%), and in combination with preoperative embolization in 39 cases (25%) and preoperative irradiation in 2 cases (1.3%). By SM grades, microsurgical resection alone was performed in 78% of patients in SM1 and SM2, 46% in SM3, and 14% in SM4. By comparison, preoperative embolization was used in 17% of SM1 and SM2 patients and 72% of SM4. AVM obliteration was achieved in 94.1% of patients on initial DSA and 98.1% on final DSA. Incomplete AVM obliterations were treated with repeat microsurgery in 4 patients and postoperative radiosurgery in 2 patients. Reasons for incomplete obliteration on final DSA were refusal for DSA (1 patient), refusal for treatment (1 patient), and loss to follow-up (1 patient). Postoperative mrs was unchanged or improved in 79.4% of patients, and mrs increased by 1 in 18.7%. Neurological deficits of any kind at last follow-up were present in 16.1% of patients, whereas EDD and PDD rates were 11.6% and 4.5%, respectively. Median length of admission was 5 days. Other complications included 19 patients with major bleeding, 6 patients with postoperative hematomas requiring evacuation, and 2 wound infections. When analyzed according to the SM grades, there were 118 patients with SM1 and SM2. Within this group, obliteration rates on initial and final DSA were 96.6% and 99.1%, respectively. EDD occurred in 9.3% and PDD in 3.4% of SM1 and SM2 patients, whereas neurological deficits of any kind at final follow-up were present in 10.2%. Median length of stay

3 138 Stroke January 2017 Figure 1. Flow chart demonstrating the breakdown of 979 patients identified in the arteriovenous malformation (AVM) database to 155 surgical patients with unruptured brain AVM for final analysis. ICH indicates intracerebral hemorrhage; IVH, intraventricular hemorrhage; and SAH, subarachnoid hemorrhage. was 4 days. By comparison, in SM3 and SM4 patients, EDD rates were 20% and 28.6%, respectively, whereas PDD rates were 10% and 0%, respectively. In summary, a greater proportion of SM3 and SM4 patients had preoperative embolization, longer length of stay, and higher EDD and PDD when compared with SM1 and SM2. Univariate and multivariate analyses were performed to identify predictive factors for poor outcome. On univariate analysis (Table 4), several covariates were significantly associated with permanent neurological deficits (P<0.05): high-grade AVM, eloquence, deep venous drainage, AVMassociated aneurysms, surgery alone, and major bleeding. On multivariate analysis, a significant interaction was seen between surgery alone and major bleeding, indicating that an adjunct procedure (embolization) in addition to surgical intervention significantly reduced the effect of major bleeding. In patients who underwent surgery alone, major bleeding was the only significant independent predictor (P<0.001) of permanent neurological deficit. For EDD, both high-grade AVM and major bleeding reached significance on univariate analysis, but only major bleeding was significant on multivariate (P<0.001). Similarly, major bleeding was the only significant factor on univariate analysis for PDD. Discussion ARUBA is the only prospective randomized trial comparing medical management with interventional treatment for unruptured bavms with a statistically significant morbidity associated with interventional treatment. 6,7 It reflects the current management of ubavms internationally as 39 active centers across 9 countries were involved. ARUBA provided prospective information on the natural history of ubavms, confirming that it is not benign because 10.1% of patients suffered a symptomatic stroke or death within 33 months. However, ARUBA was criticized because of its low enrollment rate, small sample size, short follow-up, high rate of adverse outcomes, under-representation of surgical treatment, and lack of treatment stratification Its primary outcome after interventional treatment was higher than in previous cohorts treated with microsurgery or radiosurgery without a clear explanation. It combined 3 different interventions into a single amorphous category and did not discriminate which modality was harmful. 21 ARUBA also assumed equipoise among all ubavms between medical and interventional treatments, 30 an assumption not shared by many physicians, which may explain the small proportion of patients enrolled (226/1740) By not enrolling patients, physicians have potentially introduced bias, leading to the current quandary that ARUBA tried to avoid. It raises questions about the generalizability of ARUBA data to individual ubavm treatment. AVMs are heterogeneous, with varying angioarchitecture, anatomic locations, and natural history. 13 UbAVMs considered low grade for treatment are not associated with a more benign natural history. The risks of treatment are not homogeneous where treatment strategies depend on characteristics of

4 Wong et al Microsurgery for Unruptured Brain AVMs 139 Table 1. Baseline Patient Characteristics for 155 Unruptured Brain Arteriovenous Malformation Patients Treated With Microsurgery, Including Mean Follow-Up, Clinical Presentation, Past History, and Preoperative mrs Patient Characteristics Total No. of patients 155 Age, mean (SD), y 38.5 Female (%) 88 (56.8) Follow-up, mean, mo 36.1 Clinical presentation (%) Seizure 74 (47.7) Neurological deficit 13 (8.4) Bruit 0 (0) Headache 45 (29) Asymptomatic 68 (43.9) Past history (%) Smoking 85 (54.8) Stroke/TIA 13 (8.4) Mild head injury 4 (2.6) HHT 7 (4.5) Preoperative mrs (%) 0 7 (4.5) 1 68 (43.9) 2 77 (49.7) 3 3 (1.9) HHT indicates heriditary hemorrhagic telangiectasia; mrs, modified Rankin Scale (score); and TIA, transient ischemic attack. the lesion, availability of individual treatment modalities, and institutional expertise. 10,12,17,31 34 Several retrospective series have recently been published on ubavms. The SIVMS (Scottish Intracranial Vascular Malformation Study) reported on ubavms comparing conservative management with intervention, which also found superiority in the conservative group for death or handicap at 4 years and focal neurological deficit or death at 12 years. 1 Both SIVMS and ARUBA cohorts had predominantly nonsurgical interventions and bavm obliteration rates of 63% to 71% were observed in SIVMS. 16 In contrast, Bervini et al 35 presented a 25-year microsurgical series from Sydney, Australia, of 427 unruptured bavms patients, who were stratified according to the Spetzler Ponce class 4 : for class A (n=190), the rate of permanent neurological deficit with mrs increase >1 was 1.6% (95% confidence interval, %); class B (n=107), 14.0% (95% confidence interval, %); and class C (n=44), 38.6% (95% confidence interval, %). 35,36 Subsequent sensitivity analyses on the same database by Korja et al 37 showed no statistical difference to the combined classes A and B adverse outcome rate of 7.7% when nonoperated patients were assumed to have adverse outcomes. Outcomes in ARUBA-eligible patients at University of California, San Francisco (UCSF) have also been reported. 15 Table 2. Baseline AVM Characteristics of the Surgical Cohort, Including AVM Laterality, Location, Associated Aneurysms, Overall, and Individual Components of SM Grading AVM Characteristics Total Single/multiple lesions (%) Single 152 (98.1) Multiple 3 (1.9) AVM side (% of total lesions=159) Right 84 (52.8) Left 75 (47.2) AVM location (% of total lesions=159) Frontal 47 (29.6) Parietal 44 (27.7) Temporal 40 (25.2) Occipital 16 (10.1) Insular 6 (3.8) Cerebellum/posterior fossa 5 (3.2) Periventricular 1 (0.6) Aneurysms (%) AVM associated 14 (9.0) Non-AVM associated 10 (6.5) SM grade (%) 1 52 (33.5) 2 66 (42.6) 3 30 (19.4) 4 7 (4.5) 5 0 (0) Size (%) <3 cm 112 (72.3) 3 6 cm 43 (27.7) >6 cm 0 (0) Eloquent location (%) 62 (40) Deep venous drainage (%) 42 (27.1) AVM indicates arteriovenous malformation; and SM, Spetzler Martin. Of 74 ARUBA-eligible patients, 61 patients had received intervention, which included surgical resection in 70.5% of patients (with and without preoperative embolization), and radiosurgery in the remainder. The risk of stroke or death was 14.7% in the entire intervention group when compared with 7.7% in the conservative group, though surgery still had a lower risk of stroke or death (11.6%) than the overall intervention group. Functional outcomes demonstrated no significant differences and complete obliteration of AVMs was achieved in 93% of treated patients. Similarly, Nerva et al 38 reviewed outcomes of 105 unruptured bavm patients, including a subgroup analysis of 61 ARUBA-eligible patients, on whom microsurgery was used in 61% of low-grade bavm (SM1 and SM2) and 40% of high-grade AVM. Complete obliteration was achieved in

5 140 Stroke January 2017 Figure 2. Distribution of functional outcomes (mrs) for all SM grades (A and B) and for Spetzler Martin (SM) grades 1 and 2 (C and D) at the indicated assessment intervals: initial assessment, 7 d postoperative and at last follow-up. Green: mrs unchanged or improved; orange: mrs 2 and worse than initial assessment; red: mrs 3. mrs indicates modified Rankin Scale (score). all patients. Persistent neurological deficits and mrs 3 were found in 16% and 7%, respectively, of the ARUBA-eligible cohort. In SM1 and SM2 patients, lower rates of persistent deficits and mrs 3 (7% and 0%, respectively) were reported. Interestingly, a recent multicenter retrospective radiosurgical series from 7 international institutions in 509 ARUBA-eligible patients 39 (mean follow-up 86 months) showed an AVM obliteration rate of 75%, with a postradiosurgery latency hemorrhage rate of 0.9% per year. The adverse neurological outcome, permanent neurological morbidity, and mortality rates were 13%, 5%, and 4%, respectively, which were more favorable than the primary outcome in ARUBA interventional group. Our analysis of 155 consecutive ubavm patients who were generally eligible for ARUBA and treated with microsurgery at a single multidisciplinary institution represents 29.3% of the entire ubavm cohort (528 patients) treated during the same period. The analyzed cohort had similar patient characteristics to ARUBA except for 8 patients aged <18 years and 3 patients with initial mrs >2. The SM grade distribution was also similar with no SM grade 5 patients and the majority (76%) were SM1 and SM2. Within this subset, microsurgical resection alone was performed in 78% of cases, when compared with 14% in SM4. In high-grade ubavms, preoperative embolization and neurological deficits were more common, median length of stay and follow-up were longer, and AVM obliteration rates were also lower. Outcomes from our series demonstrated a similar result to the aforementioned surgical series. 15,38,40 In our series, the overall EDD and PDD (mrs 3) rates were 12.3% and 4.5%, respectively. Postoperative DSA confirmed complete resection in 98.1% of cases, accounting for the refusal of 1 patient to undergo DSA and thus assuming persistence of AVM in that individual. In contrast, the low-grade ubavms (SM1 and SM2) in our series were associated with higher obliteration rates (99.2%) and lower PDD (3.4%), which compares favorably to the overall morbidity (2.2%) and complete obliteration rates (98.5%) presented in a recent systematic review of microsurgery on low-grade AVMs by Potts et al. 40 The mean

6 Wong et al Microsurgery for Unruptured Brain AVMs 141 Table 3. Distribution of Treatment Modalities, Outcomes, and Complications When Stratified According to SM Grade Overall, n=155 SM1, n=52 SM2, n=66 SM3, n=30 SM4, n=7 SM1 and SM2, n=118 Treatment modalities (%) Surgery only: single operation 107 (69.0) 47 (90.4) 45 (68.2) 14 (46.7) 1 (14.3) 92 (78.0) Surgery only: multiple operations 4 (2.6) 0 (0) 4 (6.1) 0 (0) 0 (0) 4 (3.4) Preoperative embolization+surgery 39 (25.2) 5 (9.6) 15 (22.7) 14 (46.7) 5 (71.4) 20 (16.9) Preoperative radiosurgery+surgery 2 (1.3) 0 (0) 1 (1.5) 1 (3.3) 0 (0) 1 (0.8) Surgery+postoperative radiosurgery 2 (1.3) 0 (0) 1 (1.5) 1 (3.3) 0 (0) 1 (0.8) Embolization+surgery+radiosurgery 1 (0.6) 0 (0) 0 (0) 0 (0) 1 (14.3) 0 (0) Outcomes (%) Cure on initial postopeartive DSA* 146 (94.2) 51 (98.1) 63 (95.5) 27 (90.0) 5 (71.4) 114 (96.6) Cure on final postoperative DSA* 152 (98.1) 52 (100) 65 (98.5) 29 (96.7) 6 (85.7) 117 (99.2) Length of admission, mean, d Length of admission, median, d mrs same or better at last FU 123 (79.4) 45 (86.5) 48 (72.7) 26 (86.7) 3 (43.9) 93 (78.8) mrs worse at last FU 32 (20.6) 7 (13.5) 18 (27.3) 4 (13.3) 4 (57.1) 25 (21.2) Complications (%) Early neurological deficit, within 7 d 47 (30.3) 11 (21.2) 18 (27.3) 13 (43.3) 5 (71.4) 29 (24.6) Early disabling deficit, mrs 3 within 7 d 19 (12.3) 5 (9.6) 6 (9.1) 6 (20.0) 2 (28.6) 11 (9.3) Permanent neurological deficit, at last FU 25 (16.1) 2 (3.8) 10 (15.2) 9 (30.0) 4 (57.1) 12 (10.2) Permanent disabling deficit, mrs 3 at last FU 7 (4.5) 0 (0) 4 (6.1) 3 (10.0) 0 (0) 4 (3.4) Postoperative hemorrhage 6 (3.9) 2 (3.8) 0 (0) 2 (6.7) 2 (28.6) 2 (1.7) Treatment modalities consisted of microsurgery with or without adjuvant embolization or radiosurgery. Outcomes included radiological and clinical outcomes. DSA indicates digital subtraction angiography; FU, follow-up; mrs, modified Rankin Scale (score); and SM, Spetzler Martin. *DSA refused by 3 patients. and median duration of follow-up in our series were 36.1 and 18 months, respectively, which is comparable to the mean follow-up in ARUBA (33 months). The short duration of follow-up in this series can be explained by our institutional practice to discharge patients from further clinical follow-up beyond 12 months postoperatively once curative resection is confirmed by DSA. In summary, the current series and other retrospective surgical cohorts have demonstrated that microsurgical resection can be performed safely and effectively in a subset of predominantly low-grade ubavms (SM1 and SM2). By definition, one can infer that such ubavms would not have >1 high-risk feature: eloquence, size >3 cm, or deep venous drainage. To identify specific prognostic factors and AVM features associated with treatment risk, univariate and multivariate analyses were performed in this study. On the basis of logistic regression, the most consistent significant predictors of permanent neurological deficit, EDD, and PDD on univariate analysis were high-grade AVM and major bleeding. Individual components of the SM grading, such as eloquence and deep drainage, were also significant predictors of permanent deficit on univariate analysis but did not reach significance in multivariate modeling. Major bleeding was the most significant predictive factor for permanent neurological deficit, EDD, and PDD on multivariate analysis. Although the threshold defined herein for major bleeding in surgical patients (>1000 ml of blood loss, or transfusion requirement of 2 U of whole blood or packed red cells) 5 is arbitrary, this may be a surrogate measure of either high-risk features in bavm specified or unspecified by the SM grade or the surgeon s technical expertise. The relevance of major bleeding as a significant predictor for poor outcome may be applicable to the role of preoperative embolization. Embolization may be used as an adjunct to reduce intraoperative bleeding by addressing high-risk features, such as an intranidal aneurysm, or to generally reduce flow through the nidus preoperatively, or to control deep arterial feeders which may be difficult to access surgically. However, risks of embolization will vary depending on the treatment strategy. Occasionally, embolization has been used for definitive cure, with reported cure rates of 15% to 50% and higher morbidity rates. 11,14,20,40 In ARUBA, embolization alone constituted 26% of cases, but the treatment strategy or specific agents were not mentioned. 11 Within the subset of low-grade AVMs, a review of endovascular series using Onyx embolization demonstrated morbidity and cure rates of 6.2% and 29.5%, respectively. 40 Although there is no established treatment algorithm or paradigm at our institution, consensus decision at the multidisciplinary conference would determine the most appropriate modality of treatment for individual bavms. In general, microsurgery would be recommended if the bavm is accessible and can be resected with minimal acceptable morbidity, usually in the setting of SM1 and SM2 and possibly SM3.

7 142 Stroke January 2017 Table 4. Univariate and Multivariate Analyses for Predictive Factors of Permanent Neurological Deficit, EDD, and PDD OR (95% CI) Significance Permanent neurological deficit Univariate analysis SM grade (high grade vs low grade) 4.78 ( ) <0.001 Eloquence 0.25 ( ) Size of nidus 0.51 ( ) 0.14 Venous drainage 0.27 ( ) Age 0.98 ( ) 0.19 AVM-associated aneurysms 3.36 ( ) Nonassociated aneurysms 0.56 ( ) 0.59 Major bleeding 6.75 ( ) <0.001 Surgery alone 0.34 ( ) Cure on initial DSA 0.28 ( ) 0.09 Multivariate analysis* SM grade (high grade vs low grade) 2.79 ( ) 0.05 AVM-associated aneurysms 2.48 ( ) 0.19 Surgery alone 0.45 ( ) 0.12 Major bleeding 3.82 ( ) Two-way interactions Surgery alone Major bleeding ( ) Surgery alone, n=107 SM grade 2.91 ( ) 0.21 AVM-associated aneurysms 1.90 ( ) 0.52 Major bleeding ( ) <0.001 Other treatments, n=48 SM grade 2.68 ( ) 0.16 AVM-associated aneurysms 1.21 ( ) 0.88 Major bleeding 0.62 ( ) 0.61 EDD Univariate analysis SM grade (high grade vs low grade) 2.98 ( ) Age 1.02 ( ) 0.45 AVM-associated aneurysms 2.29 ( ) 0.24 Nonassociated aneurysms 0.83 ( ) 0.87 Major bleeding 9.16 ( ) <0.001 Surgery alone 0.52 ( ) 0.19 Cure on initial DSA 0.37 ( ) 0.25 Multivariate analysis* SM grade 0.17 ( ) 0.36 (Continued ) Table 4. Continued OR (95% CI) Significance Major bleeding 7.56 ( ) <0.001 PDD Univariate analysis SM grade (high grade vs low grade) 2.51 ( ) 0.24 Age 1.02 ( ) 0.61 AVM-associated aneurysms 1.73 ( ) 0.62 Nonassociated aneurysms 1.70 (0.001 ) 0.99 Major bleeding 6.19 ( ) 0.02 Surgery alone 0.58 ( ) 0.49 Cure on initial DSA 5.82 (0 ) 0.99 Significant 2-way interaction between surgery alone and major bleeding for PDD (P=0.002); separate models for surgery alone and other treatments. AVM indicates arteriovenous malformation; CI, confidence interval; DSA, digital subtraction angiography; EDD, early disabling deficit; OR, odds ratio; PDD, permanent disabling deficit; and SM, Spetzler Martin. *Models drawn on significant univariate variables (P<0.05). This can be observed in the distribution of SM grading in our cohort, with few SM4 or SM5 patients treated. Adjunctive preoperative embolization may be used to aid microsurgical resection by addressing high-risk features, such as intranidal aneurysms, or major arteriovenous shunts, or the deep nidal component of a multicompartment AVM, and is more commonly used in higher SM grades. For deeply located bavm, stereotactic radiosurgery may be recommended. In summary, not all bavm should be considered homogeneously for intervention, and intervention should be individualized according to the AVM characteristics and the expertise of the treating institution. The current series highlights the need for continuing investigation into identifying prognostic factors to appropriately select ubavm patients for treatment. Potts et al 40 recently proposed a new randomized trial for low-grade bavm to ultimately address concerns raised by ARUBA, BARBADOS trial (Beyond ARUBA: Randomized Low-Grade Brain AVM Study: Observation Versus Surgery). It is hoped that this would inspire greater participation from neurosurgeons in high-volume centers to produce a positive result against ARUBA. However, similar issues with funding and duration of follow-up may also apply to BARBADOS, as has plagued ARUBA previously. Given the scarcity of AVMs and the complexity of their treatments, the cumulative experience from high-volume surgical centers on bavm, 3 with this study included, will supply data to refine future prospective intervention trials for all ubavms, such that an evidence-based treatment algorithm may be formulated. Limitations of Current Study Although our data were collected prospectively, it remains a retrospective review from a single institution. Surgical patients were selected mainly based on the SM grade, angioarchitecture, and location of bavm, and thus, low-grade lesions were heavily represented. Outcomes were assessed by clinicians who were not blinded or independent of the care of the

8 Wong et al Microsurgery for Unruptured Brain AVMs 143 patients. Adjuvant endovascular techniques have evolved in the past 20 years, the results, therefore, reflect past treatment trends, rather than the latest techniques. 17 Conclusions Although ARUBA remains the only level 1 evidence for management of ubavms, our study demonstrates the safety of microsurgical resection in a carefully selected population of ubavm patients based on SM grade, AVM angioarchitecture, and judicious use of preoperative embolization. It adds weight to the emerging evidence from other microsurgical series consisting of predominantly low-grade ubavms (SM1 and SM2) that microsurgery can be efficacious and challenges the conclusion that medical management is superior to all interventions. Data from retrospective surgical series, such as this study, may be used to develop future prospective trials and a potential treatment algorithm for ubavms. Acknowledgments All authors contributed to the data collection, writing, statistical analysis, and illustration of the article. Dr Tymianski is a Canada Research Chair (tier 1) in translational stroke research. Sources of Funding This work was funded by the aneurysm research fund, Neurovascular Therapeutics Program, University Health Network. Dr Wong was supported by the Royal Australasian College of Surgeons for the Stuart-Morson Travel Scholarship. None. Disclosures References 1. Al-Shahi Salman R, White PM, Counsell CE, du Plessis J, van Beijnum J, Josephson CB, et al; Scottish Audit of Intracranial Vascular Malformations Collaborators. Outcome after conservative management or intervention for unruptured brain arteriovenous malformations. JAMA. 2014;311: doi: /jama Mohr JP, Parides MK, Stapf C, Moquete E, Moy CS, Overbey JR, et al; International ARUBA Investigators. Medical management with or without interventional therapy for unruptured brain arteriovenous malformations (ARUBA): a multicentre, non-blinded, randomised trial. Lancet. 2014;383: doi: /S (13) Kim H, Abla AA, Nelson J, McCulloch CE, Bervini D, Morgan MK, et al. Validation of the supplemented Spetzler-Martin grading system for brain arteriovenous malformations in a multicenter cohort of 1009 surgical patients. Neurosurgery. 2015;76:25 31, discussion 31. doi: / NEU Spetzler RF, Ponce FA. A 3-tier classification of cerebral arteriovenous malformations. Clinical article. J Neurosurg. 2011;114: doi: / JNS Schulman S, Angerås U, Bergqvist D, Eriksson B, Lassen MR, Fisher W; Subcommittee on Control of Anticoagulation of the Scientific and Standardization Committee of the International Society on Thrombosis and Haemostasis. Definition of major bleeding in clinical investigations of antihemostatic medicinal products in surgical patients. J Thromb Haemost. 2010;8: doi: /j x. 6. Mohr JP. Results of ARUBA are applicable to most patients with nonruptured arteriovenous malformations. Stroke. 2014;45: doi: /STROKEAHA Molina CA, Selim MH. Unruptured brain arteriovenous malformations: keep calm or dance in a minefield. Stroke. 2014;45: doi: /STROKEAHA Wong J, Radovanovic I, Tymianski M. The impact of ARUBA on the management of unruptured brain arteriovenous malformations: review of literature. Jpn J Neurosurg. 2015;24: doi: /jcns Cockroft KM, Jayaraman MV, Amin-Hanjani S, Derdeyn CP, McDougall CG, Wilson JA. A perfect storm: how a randomized trial of unruptured brain arteriovenous malformations (ARUBA s) trial design challenges notions of external validity. Stroke. 2012;43: doi: / STROKEAHA Mathiesen T. Arguments against the proposed randomised trial (ARUBA). Neuroradiology. 2008;50: doi: /s Mocco J, O Kelly C, Arthur A, Meyers PM, Hirsch JA, Woo HH, et al. Randomized clinical trials: the double edged sword. J Neurointerv Surg. 2013;5: doi: /neurintsurg Amin-Hanjani S. ARUBA results are not applicable to all patients with arteriovenous malformation. Stroke. 2014;45: doi: / STROKEAHA Bambakidis NC, Cockroft K, Connolly ES, Amin-Hanjani S, Morcos J, Meyers PM, et al. Preliminary results of the ARUBA study. Neurosurgery. 2013;73:E379 E381. doi: /NEU Knopman J, Stieg PE. Management of unruptured brain arteriovenous malformations. Lancet. 2014;383: doi: / S (14) Rutledge WC, Abla AA, Nelson J, Halbach VV, Kim H, Lawton MT. Treatment and outcomes of ARUBA-eligible patients with unruptured brain arteriovenous malformations at a single institution. Neurosurg Focus. 2014;37:E8. doi: / FOCUS Starke RM, Sheehan JP, Ding D, Liu KC, Kondziolka D, Crowley RW, et al. Conservative management or intervention for unruptured brain arteriovenous malformations. World Neurosurg. 2014;82:e668 e669. doi: /j.wneu Conger A, Kulwin C, Lawton MT, Cohen-Gadol AA. Endovascular and microsurgical treatment of cerebral arteriovenous malformations: current recommendations. Surg Neurol Int. 2015;6:39. doi: / Pollock BE, Link MJ, Brown RD. The risk of stroke or clinical impairment after stereotactic radiosurgery for ARUBA-eligible patients. Stroke. 2013;44: doi: /STROKEAHA Ross J, Al-Shahi Salman R. Interventions for treating brain arteriovenous malformations in adults. Cochrane Database Syst Rev. 2010;7:CD doi: / CD pub Russin J, Spetzler R. Commentary: the ARUBA trial. Neurosurgery. 2014;75:E96 E97. doi: /NEU Grasso G. The ARUBA study: what is the evidence? World Neurosurg. 2014;82:e576. doi: /j.wneu Lawton MT, Abla AA. Management of brain arteriovenous malformations. Lancet. 2014;383: doi: /S (14) Davidson AS, Morgan MK. How safe is arteriovenous malformation surgery? A prospective, observational study of surgery as firstline treatment for brain arteriovenous malformations. Neurosurgery. 2010;66: , discussion 504. doi: /01.NEU Spetzler RF, Martin NA. A proposed grading system for arteriovenous malformations. J Neurosurg. 1986;65: doi: / jns van Beijnum J, van der Worp HB, Buis DR, Al-Shahi Salman R, Kappelle LJ, Rinkel GJ, et al. Treatment of brain arteriovenous malformations: a systematic review and meta-analysis. JAMA. 2011;306: doi: /jama Lawton MT; UCSF Brain Arteriovenous Malformation Study Project. Spetzler-Martin Grade III arteriovenous malformations: surgical results and a modification of the grading scale. Neurosurgery. 2003;52: , discussion 748. doi: /01.NEU C. 27. Ding D, Yen CP, Xu Z, Starke RM, Sheehan JP. Radiosurgery for patients with unruptured intracranial arteriovenous malformations. J Neurosurg. 2013;118: doi: / JNS Pollock BE, Flickinger JC, Lunsford LD, Maitz A, Kondziolka D. Factors associated with successful arteriovenous malformation radiosurgery. Neurosurgery. 1998;42: , discussion Wegner RE, Oysul K, Pollock BE, Sirin S, Kondziolka D, Niranjan A, et al. A modified radiosurgery-based arteriovenous malformation grading scale and its correlation with outcomes. Int J Radiat Oncol Biol Phys. 2011;79: doi: /j.ijrobp Davis SM, Donnan GA. Unruptured brain arteriovenous malformations: another asymptomatic conundrum. Stroke. 2007;38:3312. doi: / STROKEAHA Starke RM, Komotar RJ, Connolly ES. A randomized trial of unruptured brain arteriovenous malformations. Neurosurgery. 2013;73:N13 N15. doi: /01.neu e3.

9 144 Stroke January Heros RC. Multimodality treatment of cerebral arteriovenous malformations: modern treatment of cerebral arteriovenous malformations. World Neurosurg. 2014;82: doi: /j.wneu Pollock BE. The alchemy of brain arteriovenous malformation management. World Neurosurg. 2015;83: doi: /j. wneu Spears J, Terbrugge KG, Moosavian M, Montanera W, Willinsky RA, Wallace MC, et al. A discriminative prediction model of neurological outcome for patients undergoing surgery of brain arteriovenous malformations. Stroke. 2006;37: doi: /01. STR Bervini D, Morgan MK, Ritson EA, Heller G. Surgery for unruptured arteriovenous malformations of the brain is better than conservative management for selected cases: a prospective cohort study. J Neurosurg. 2014;121: doi: / JNS Bervini D, Morgan MK, Ritson EA, Heller G. Surgery for unruptured arteriovenous malformations of the brain is better than conservative management for selected cases: a prospective cohort study. Neuro-Chirurgie. 2014;60: Korja M, Bervini D, Assaad N, Morgan MK. Role of surgery in the management of brain arteriovenous malformations: prospective cohort study. Stroke. 2014;45: doi: /STROKEAHA Nerva JD, Mantovani A, Barber J, Kim LJ, Rockhill JK, Hallam DK, et al. Treatment outcomes of unruptured arteriovenous malformations with a subgroup analysis of ARUBA (A Randomized Trial of Unruptured Brain Arteriovenous Malformations)-eligible patients. Neurosurgery. 2015;76: , discussion 570, quiz 570. doi: / NEU Ding D, Starke RM, Kano H, Mathieu D, Huang P, Kondziolka D, et al. Radiosurgery for cerebral arteriovenous malformations in A Randomized Trial of Unruptured Brain Arteriovenous Malformations (ARUBA)- eligible patients: a multicenter study. Stroke. 2016;47: doi: /STROKEAHA Potts MB, Lau D, Abla AA, Kim H, Young WL, Lawton MT; UCSF Brain AVM Study Project. Current surgical results with low-grade brain arteriovenous malformations. J Neurosurg. 2015;122: doi: / JNS14938.